Dunte Rolle
####Run through this example and try to understand what is going on with the data
####So, lets load RMarkdown
library(rmarkdown)set messages to FALSE on everything (prevents certain boring things from being shown in the results)
knitr::opts_chunk$set(echo = TRUE, message=FALSE,warning=FALSE,collapse = TRUE)PACKAGES
library(reshape2)
library(ggplot2)
library(dplyr)
library(plotly)
library(viridis)
library(data.table)
library(pheatmap)
library(tidyverse)
library(ggthemes)
library(clipr)
library(tidyr)
library(Rcpp)COLORS
mycolors<-c(viridis(15))
felix_cols<-mycolors[c(5,2)]
felix_4cols<-mycolors[c(15,10,8,2)]
plain_cols1<-c("pink","gray")
plain_cols2<-c("purple","gray")
pats_cols<-colorRampPalette(c("#FDE725FF", "white","red"))(21)
leos_cols<-colorRampPalette(c("white","blue"))(10)
LOAD DATA AND MAKE AN OVERALL HEATMAP
## load the dataset
breast_cancer_cells<- read_csv(file="breast_cancer_cells.csv")
#and then view it
view(breast_cancer_cells)
now let’s visualize the dataset and look for initial trends. We can do this by making a matrix so and then a heatmap to visualize the data
#make a matrix of just raw the data
m1br_can<-breast_cancer_cells %>%select(MCF10A_1:SKBR3_2) %>% as.matrix() %>% round(.,2)
# Accidentally deleted '>' stopping code from knitting
# what does that look like?
head(m1br_can)
## MCF10A_1 MCF10A_2 MCF7_1 MCF7_2 MDA231_1 MDA231_2 MDA468_1 MDA468_2
## [1,] 9.54 4.58 5.07 5.42 25.43 27.42 4.56 3.88
## [2,] 14.00 11.58 6.49 6.64 9.80 10.31 6.84 8.75
## [3,] 10.22 8.29 11.55 10.82 12.48 10.11 9.54 8.46
## [4,] 9.00 6.35 13.40 14.82 14.94 11.33 6.87 7.92
## [5,] 8.21 4.44 12.08 9.82 16.51 12.34 15.43 8.50
## [6,] 12.51 15.84 8.05 8.38 6.78 8.46 4.48 7.18
## SKBR3_1 SKBR3_2
## [1,] 8.46 5.64
## [2,] 11.91 13.67
## [3,] 9.10 9.43
## [4,] 8.54 6.82
## [5,] 7.93 4.75
## [6,] 13.27 15.05## make a heatmap from the data
pheatmap(m1br_can, color=pats_cols,cellwidth=30,cellheight=.03,cluster_cols=FALSE,cluster_rows=TRUE,legend=TRUE,fontsize = 7,scale="column")
##INTERPRETATION##
## What can you see in this figure? are the repeated measures/reps similar or different? What does this say about the precision and accuracy of them?
##How does the control compare to the variables? Is this what you might expect? Why? What would you look for in the literature to support this idea?
#looks like _______ line is different than the other cells lines. We will take a closer look at that in the next section, once we've done some further data manipulations
Data Manipulation
## first, make new columns that combine the two reps for each variable by averaging them
breast_cancer_cells2<-breast_cancer_cells %>% mutate(
mean_MCF10A= ((MCF10A_1+ MCF10A_2)/2),
mean_MCF7= ((MCF7_1 + MCF7_2)/2),
mean_MDA231= ((MDA231_1 + MDA231_2)/2),
mean_MDA468= ((MDA468_1 + MDA468_2)/2),
mean_SKBR3= ((SKBR3_1 + SKBR3_2)/2))
view(breast_cancer_cells2)
breast_cancer_cells2<-breast_cancer_cells2 %>% mutate(
log_MCF7=log2(mean_MCF7/mean_MCF10A),
log_MDA231=log2(mean_MDA231/mean_MCF10A),
log_MDA468=log2(mean_MDA468/mean_MCF10A),
log_SKBR3=log2(mean_SKBR3/mean_MCF10A))
colnames(breast_cancer_cells2)
## [1] "Gene_Symbol" "Description"
## [3] "Peptides" "MCF10A_1"
## [5] "MCF10A_2" "MCF7_1"
## [7] "MCF7_2" "MDA231_1"
## [9] "MDA231_2" "MDA468_1"
## [11] "MDA468_2" "SKBR3_1"
## [13] "SKBR3_2" "pvalue_MCF7_vs_MCF10A"
## [15] "pvalue_MDA231_vs_MCF10A" "pvalue_MDA468_vs_MCF10A"
## [17] "pvalue_SKBR3_vs_MCF10A" "mean_MCF10A"
## [19] "mean_MCF7" "mean_MDA231"
## [21] "mean_MDA468" "mean_SKBR3"
## [23] "log_MCF7" "log_MDA231"
## [25] "log_MDA468" "log_SKBR3"
#did it create your new columns?
## then make some new columns that store the log ratios of the variable means you just created, as compared to the control
m2br_can<-breast_cancer_cells2 %>% select(log_MCF7:log_SKBR3) %>% as.matrix() %>% round(.,2)
pheatmap(m2br_can, color=pats_cols,cellwidth=30,cellheight=.03,cluster_cols=FALSE,cluster_rows=TRUE,legend=TRUE,fontsize = 7,scale="column")
#did it create your new columns?## ALTERNATIVE CHOICE: make a matrix of just log values the data and a heat map of that. How do the two heat maps compare? This is not possible if your values contain 0's
Diving deeping with VOLCANO PLOTS
##BASIC VOLCANO PLOT
## Add a column that stores the negative log of the pvalue of interest (in this case, SKBR3)
breast_cancer_cells2<-breast_cancer_cells2 %>% mutate(neglog_SKBR3 = -log10(pvalue_SKBR3_vs_MCF10A))
## Use ggplot to plot the log ratio of SKBR3 against the -log p-value pvalue_SKBR3_vs_MFCA10A
volcano_plot<-breast_cancer_cells2 %>% ggplot(aes(x = log_SKBR3, y = neglog_SKBR3, description = Gene_Symbol)) +
geom_point(alpha = 0.7, color = "skyblue")
#to view it, type:
volcano_plot
## BETTER VOLCANO PLOT
## Define the significant ones (by using ifelse and setting # parameters) so they can be colored
breast_cancer_cells2<-breast_cancer_cells2 %>% mutate(significance=ifelse((log_SKBR3>2 & neglog_SKBR3>2.99),"UP", ifelse((log_SKBR3<c(-2) & neglog_SKBR3>2.99),"DOWN","NOT SIG")))
## Some standard colors
## volcano plot as before with some added things
plain_cols3<-c("red","gray","blue")
better_volcano_plot <- breast_cancer_cells2 %>%
ggplot(aes(x=log_SKBR3,y=neglog_SKBR3,description=Gene_Symbol,color=significance))+
geom_point(alpha=0.7)+
scale_color_manual(values=plain_cols3)+
xlim(-6,6)+
theme_bw()+
theme(axis.text = element_text(colour = "black",size=14))+
theme(text = element_text(size=14))+
labs(x="log ratio SKBR3 vs MCF10A", y = "-log(p-value)")
#to view it, type
better_volcano_plot
#check viewer and / or plots to see it
## use ggplotly to see hover over the points to see the gene names. Record these for the next step
ggplotly(better_volcano_plot)
##INTERPRETATION##
##
#Why? How?Barplots of significant points of interest
EXAMPLES OF A COUPLE PROTEINS or GENES
# Make a pivot longer table of the breast_cancer_cells data for MCF10A_1:SKBR3_2
breast_cancer_cells_long<-pivot_longer(breast_cancer_cells2, cols = c(MCF10A_1:SKBR3_2), names_to = 'variable')%>% select(-c(pvalue_MCF7_vs_MCF10A:pvalue_SKBR3_vs_MCF10A))
head(breast_cancer_cells_long)
## # A tibble: 6 × 16
## Gene_Symbol Description Peptides mean_MCF10A mean_MCF7 mean_MDA231 mean_MDA468
## <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 NES Nestin OS=… 7 7.06 5.24 26.4 4.22
## 2 NES Nestin OS=… 7 7.06 5.24 26.4 4.22
## 3 NES Nestin OS=… 7 7.06 5.24 26.4 4.22
## 4 NES Nestin OS=… 7 7.06 5.24 26.4 4.22
## 5 NES Nestin OS=… 7 7.06 5.24 26.4 4.22
## 6 NES Nestin OS=… 7 7.06 5.24 26.4 4.22
## # ℹ 9 more variables: mean_SKBR3 <dbl>, log_MCF7 <dbl>, log_MDA231 <dbl>,
## # log_MDA468 <dbl>, log_SKBR3 <dbl>, neglog_SKBR3 <dbl>, significance <chr>,
## # variable <chr>, value <dbl>
## based on the gene symbols from plotly, select a few proteins from the table. Select just the data and gene symbols and pivot longer
breast_cancer_Down<-breast_cancer_cells_long
breast_cancer_Down<-breast_cancer_cells_long%>% filter(Gene_Symbol=="APOA1" | Gene_Symbol=="HLA-A" | Gene_Symbol=="MYO1B" | Gene_Symbol=="HMGN5")
head(breast_cancer_Down)
## # A tibble: 6 × 16
## Gene_Symbol Description Peptides mean_MCF10A mean_MCF7 mean_MDA231 mean_MDA468
## <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 HLA-A HLA class … 3 6.02 2.46 4.39 33.2
## 2 HLA-A HLA class … 3 6.02 2.46 4.39 33.2
## 3 HLA-A HLA class … 3 6.02 2.46 4.39 33.2
## 4 HLA-A HLA class … 3 6.02 2.46 4.39 33.2
## 5 HLA-A HLA class … 3 6.02 2.46 4.39 33.2
## 6 HLA-A HLA class … 3 6.02 2.46 4.39 33.2
## # ℹ 9 more variables: mean_SKBR3 <dbl>, log_MCF7 <dbl>, log_MDA231 <dbl>,
## # log_MDA468 <dbl>, log_SKBR3 <dbl>, neglog_SKBR3 <dbl>, significance <chr>,
## # variable <chr>, value <dbl>
## make barplots facetted by Gene Symbol (when working with other data sets - change x=order to x = variable)
breast_cancer_plot_Down<-breast_cancer_Down
breast_cancer_plot_Down<-breast_cancer_Down%>%ggplot(aes(x=variable,y=value))+
geom_bar(stat="identity",fill="purple")+
facet_wrap(~Gene_Symbol)+
theme_bw()+
theme(axis.text = element_text(colour = "black",size=10))+
theme(text = element_text(size=14))+
theme(axis.text.x = element_text(angle=45, hjust=1))+
labs(x="sample",y="relative intensity")
breast_cancer_plot_Down
#check viewer and / or plots to see it
##INTERPRETATION## ## What can you see in this figure? are the repeated measures/reps similar or different? What does this say about the precision and accuracy of them? ##How does the control compare to the variables? Is this what you might expect? Why? What would you look for in the literature to support this idea?
## Same process for the upregulated ones
breast_cancer_UP<-breast_cancer_cells_long
breast_cancer_UP<-breast_cancer_cells_long%>% filter(Gene_Symbol=="TDP2" | Gene_Symbol=="FNBP1L" | Gene_Symbol=="KRT23" | Gene_Symbol=="GCLC" | Gene_Symbol=="DENND4C")
head(breast_cancer_UP)
## # A tibble: 6 × 16
## Gene_Symbol Description Peptides mean_MCF10A mean_MCF7 mean_MDA231 mean_MDA468
## <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 TDP2 Isoform 2 … 3 2.04 5.3 6.66 23.3
## 2 TDP2 Isoform 2 … 3 2.04 5.3 6.66 23.3
## 3 TDP2 Isoform 2 … 3 2.04 5.3 6.66 23.3
## 4 TDP2 Isoform 2 … 3 2.04 5.3 6.66 23.3
## 5 TDP2 Isoform 2 … 3 2.04 5.3 6.66 23.3
## 6 TDP2 Isoform 2 … 3 2.04 5.3 6.66 23.3
## # ℹ 9 more variables: mean_SKBR3 <dbl>, log_MCF7 <dbl>, log_MDA231 <dbl>,
## # log_MDA468 <dbl>, log_SKBR3 <dbl>, neglog_SKBR3 <dbl>, significance <chr>,
## # variable <chr>, value <dbl>
breast_cancer_plot_UP<-breast_cancer_UP
breast_cancer_plot_UP<-breast_cancer_UP%>% ggplot(aes(x=variable,y=value))+
geom_bar(stat="identity",fill="yellow")+
facet_wrap(~Gene_Symbol)+
theme_bw()+
theme(axis.text = element_text(colour = "black",size=10))+
theme(text = element_text(size=14))+
theme(axis.text.x = element_text(angle=45, hjust=1))+
labs(x="sample",y="relative intensity")
breast_cancer_plot_UP
#check viewer and / or plots to see it
##INTERPRETATION## ## What can you see in this figure? are the repeated measures/reps similar or different? What does this say about the precision and accuracy of them? ##How does the control compare to the variables? Is this what you might expect? Why? What would you look for in the literature to support this idea?
#interpretation HINT:insert a chunk and create two seprate lines of code that filter for your specific upregulated genes/proteins of interest and selects for only their gene symbols and descriptions. Do this for the downregulated as well. This will generate two list of the descriptors for each gene of interest, helping you understand your figures. Be sure to view it, not just ask for the head of the table generated.
WRAP UP
##now you can knit this and publish to save and share your code. Use this to work with either the brain or breast cells and the Part_C_template to complete your lab 6 ELN. ##Annotate when you have trouble and reference which line of code you need help on ## good luck and have fun!